The I Understand Understand You + (IUUY+) series which is a continuously evolving framework. It is a decentralized, cognitive AI framework that moves beyond literal translation to interpret the intent, tone, and cultural nuance of global English dialects. Other existing versions of this Applied Cognitive Sociolinguistics within Collaborative Computing (CSCW) system include the Intra-English Dialect Interpretattion (IEDI), Choose Your Character Experience (CYCx) models.
Standard speech models (like Whisper or Siri) are trained on "Standard" Western English. When they hear a dialect phrase, they transcribe the words but miss the meaning.
Standard AI Output:
"I beg, what are we going to eat?"
❌ Interpretation: The speaker is pleading for food.
We combine acoustic profiling (sound) with pragmatic reasoning (logic) to create a system that understands who is speaking and why.
Uses Wav2Vec2 to detect accent clusters (e.g., West African vs. North American) from raw audio, identifying the speaker's origin before transcription.
Dynamically loads cultural "Codebooks" (JSON) based on the detected accent. This teaches the LLM that "Rain check" is an American idiom, not a weather report.
Uses IPFS and Pure Chain to create an immutable audit trail of all dialect definitions, ensuring data sovereignty for contributing communities.
Interactive Demo Loading...
Open in Hugging Face Spaces ↗Running on Gemini 1.5 Pro (Dynamic Switching) & Whisper Large-v3.
A deep dive into the IEDI-M² ecosystem, featuring discussions on the evolution of the framework and demonstrations of previous system iterations.
Watch on YouTubeThis research not only improves active listening and fosters inclusion but also advances inclusive AI ethics in line with major international frameworks.
The IEDI framework represents a step toward inclusive, dialect-aware AI capable of interpreting rather than normalizing English dialects.